Random Set

Random sets, mathematical objects representing uncertainty over sets of possibilities, are increasingly used to model complex systems and improve machine learning robustness. Current research focuses on developing new probability distributions for random sets, particularly those capable of controlling positive and negative dependence between elements, and applying these to refine algorithms like multi-object tracking and Bayesian filtering. This work aims to improve the reliability and uncertainty quantification of machine learning models, especially in high-stakes applications where understanding a model's confidence is crucial, such as in safety-critical systems. The development of novel architectures like Random-Set Convolutional Neural Networks exemplifies this trend towards more robust and uncertainty-aware AI.